Character deformation pipeline for computer-generated animation

Abstract

A deformation system for animation abstracts the notion of per-point deformation to create a pipeline, including a number of deformation modules capable of handling animation of geometry (such as characters), dynamics (such as simulations) and / or effects (such as particle systems). Deformation pipelines are defined, that work as templates capable of deforming and animating families of similar characters. Support is provided for various binding modes, including sequential, parallel, blend, and hierarchical, so as to facilitate several techniques for combining deformations.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS[0001]This application is related to the following commonly owned and co-pending U.S. patent application, the disclosure of which is incorporated herein by reference:[0002]U.S. patent application Ser. No. 10 / 769,154, entitled “Wrap Deformation Using Subdivision Surfaces,” filed Jan. 29, 2004.BACKGROUND OF THE INVENTION[0003]1. Field of the Invention[0004]The present invention relates generally to computer-generated graphics and animation, and more particularly to techniques for defining and establishing a deformation pipeline for use in computer animation.[0005]2. Description of the Background Art[0006]A central task in computer-generated (CG) animation is the construction of efficient and flexible deformations of three-dimensional (3D) characters in a way that satisfy the artistic demands of animators. When such deformations are provided with sufficient flexibility, animators are able to bring a higher degree of expressiveness to their 3D char...

AI Technical Summary

Benefits of technology

[0032]The present invention provides a more flexible system for character deformation that allows digital artists to create more fluid deformations for their 3D characters, thus approximating the techniques available to 2D artists. According to the techniques of the present invention, a character deformation pipeline is established for CG animation, so as to address the above-described needs and to avoid the limitations of prior art schemes.

[0033]According to one embodiment, the invention is implemented using a data stream abstraction of one or more CG models. A serialized stack-like pipeline of modules is established; each module takes the data stream, deforms it, and passes it to the next module. This serialized architecture for the pipeline facilitates complex layering schemes that conventionally would require complex tree dependencies. The serialized architecture of the present invention unifies known layering schemes into a single implementation.

[0034]According to one embodiment, the present invention accepts CG models based on simple and common geometrical objects such as polygonal meshes, NURBS, B-Splines, subdivision surfaces, and the like, while supporting advanced layering deformation schemes for such objects.

Problems solved by technology

Construction of a character rig for animating models using conventional techniques tends to be labor-intensive, requiring a great deal of artistic experience in order to make the best use of the deformers available in the commercial package to drive the model in the desired manner.

Conventionally, setting up the weights for the joints in the skeleton on a given character is a manual process that can be very labor-intensive.

In general, any significant editing changes on the CG model cause the weighting work done on the skeleton binding of a character to be voided.

In addition, a character rig of a CG model cannot be easily transferred to another character that might be morphologically similar but different in terms of CG model (for example, having a different number of surfaces, number of vertices, alignment, or the like).

Accordingly, character rigs generated by conventional software packages are not generic, meaning that they cannot be easily transferred from one CG model to another.

An additional challenge that exists in animation today is to bridge the gap between traditional (2D) and CG (3D) animation.

Conventional systems are ill suited for such bridging, because of differences in character rigging style and also because of limitations in the layering of deformations (the way in which multiple deformers apply their effect onto the same surface of a CG model) offered by current commercially available and proprietary systems.

However, typical CG characters for film are more complex, having numerous secondary motions taken care of by other deformers in the rig that superimpose their effects on top of the skeleton binding.

Conventional implementations of these sequential, blending and parallel layering schemes are, in general, very diverse and difficult to unify into a single scheme.

Another significant limitation of conventional rig systems is an inability to provide a simple and model-independent way to create a hierarchical layering of deformations.

For several reasons, existing hierarchical methodologies fall short of the actual CG production needs.

Firstly, because the hierarchy is built in to the surface representation, hierarchical deformations do not generally work with surfaces such as polygonal meshes, NURBS, or Catmull-Clark surfaces.

Secondly, hierarchical surfaces such as HB-Splines often fail to adapt satisfactorily to production changes.

HB-Splines are also heavy structures that take a great deal of time to update.

Images